With clock frequencies rapidly approaching the microwave range, maintaining signal integrity and controlling electromagnetic interference in electronic (digital) systems is becoming increasingly difficult. Crosstalk through mutual trace inductances and capacitors, ground bounce, clock skew, signal reflections in incorrectly terminated routes, RF radiation and pickup can no longer be ignored as in previous, relatively low frequency digital systems. The various integrated circuit (IC)-to-package and package-to-printed circuit board (PCB) interconnects must now be treated as RF transmission lines, and the characteristic line impedances of the interconnects must be matched to the signal source impedance and kept constant over the various transitions from IC, through the IC package and socket, to the PCB.
Conventional IC-to-package and package-to-PCB interconnect structures (e.g., wirebond or flip-chip structures) are difficult to shield, and exhibit impedances that are very hard to control. The resulting signal integrity problem is fundamentally ignored in commercial high pin count, mass produced ICs (e.g., memory devices, microprocessors, and linear ICs) and associated systems. As a result, the maximum operation speed of these devices is limited by signal distortions due to interconnect mismatch, rather than transistor performance. In radio frequency integrated circuits (RFIC), the number of interconnects is much smaller (a few dozen at most), but the frequency limitations of bondwires significantly impact circuit performance. Careful bondwire shaping and extensive modeling of their artifacts are expensive but common practice in the industry. However, the inherent parasitic inductance of these bondwires remains an unaddressed problem.
What is needed is an efficient and economical spring structure for IC probing or permanent IC interconnects that overcomes the signal integrity problems of conventional structures.